Despite aggressive chemotherapy and often stem cell transplant, 40% of pediatric AML patients still relapse, and most of those patients will die of refractory disease. It is well established that the microenvironment provided by stromal cells in the bone marrow niche confers a strong survival advantage to some AML cells. Recent studies have shown that AML cells recovered from bone marrow after treatment with chemotherapy are different from the untreated bulk disease population in their metabolic activities, gene expression profiles, and interactions with the microenvironment. Currently, chemoresistant AML cells cannot be distinguished from the chemosensitive bulk population at diagnosis because there is no reliable biomarker that identifies them prospectively. However, because chemoresistant cells interact differently with the microenvironment, it is likely that they have a distinct environment-induced signaling profile that can be leveraged for targeted therapeutics. The ultimate goal of our research is to identify the mechanisms by which AML cells survive chemotherapy, and by disabling those mechanisms, to achieve a meaningful advance in AML treatment. Toward that goal, the objective of this application is to identify and block the soluble factor-induced signaling pathways that specifically enable a subset of AML blasts in the bone marrow niche to survive chemotherapy. We have shown that many AML cases contain subpopulations of blasts that respond differently to the cytokines and growth factors in the bone marrow environment. For example, subpopulations with features of leukemia stem cells have more robust inducible STAT3 pathway activation, compared to bulk AML cells. Our central hypothesis is that relapse-driving AML cells respond to their microenvironment differently from chemosensitive cells. Therefore, the resistant cells can be identified by high-dimensional functional profiling methods such as mass cytometry, and they can be targeted by appropriate signaling pathway inhibitors. We will test this hypothesis with two Specific Aims. In Aim 1, we will study primary AML samples collected prior to treatment and stimulated with stroma-conditioned medium (CM) as a physiological source of bone marrow cytokines and growth factors. We will use mass cytometry to differentiate the capacity of distinct subpopulations to activate downstream pathways, and we will test the effects of inhibitors of these pathways on functions such as self-renewal in vitro and leukemia initiation in vivo. In Aim 2, we will study end of induction samples from the same patients, selected for known minimal residual disease. By mass cytometry, we will identify therapeutic vulnerabilities of chemoresistant cells by determining the CM- induced signaling profiles of the residual AML cells. We will test the efficacy of adding targeted inhibitors to chemotherapy in our in vitro and in vivo chemoresistance assays. By generating high-dimensional protein profiles of physiologically stimulated AML cells, we will deconstruct the complexity of AML to determine the signaling pathways that specifically support the survival of the cellular subset responsible for treatment failure. With this knowledge, we can devise therapies to block those pathways and eliminate the resistant cells that drive relapse.